Commit 8eb79863 authored Dec 29, 2008 by Eric W. Biederman Committed by David S. Miller Dec 29, 2008

netns: foreach_netdev_safe is insufficient in default_device_exit



During network namespace teardown we either move or delete
all of the network devices associated with a network namespace.
In the case of veth devices deleting one will also delete it's
pair device.  If both devices are in the same network namespace
then for_each_netdev_safe is insufficient as next may point
to the second veth device we have deleted.

To avoid problems I do what we do in __rtnl_kill_links and
restart the scan of the device list, after we have deleted
a device.

Currently dev_change_netnamespace does not appear to suffer from
this problem, but wireless devices are also paired and likely
should be moved between network namespaces together.  So I have
errored on the side of caution and restart the scan of the network
devices in that case as well.

Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

parent 91b208c7

net/core/dev.c

+5 −3

Original line number	Diff line number	Diff line
		@@ -5066,13 +5066,14 @@ static struct pernet_operations __net_initdata netdev_net_ops = {

		static void __net_exit default_device_exit(struct net *net)
		{
		struct net_device dev, next;
		struct net_device *dev;
		/*
		* Push all migratable of the network devices back to the
		* initial network namespace
		*/
		rtnl_lock();
		for_each_netdev_safe(net, dev, next) {
		restart:
		for_each_netdev(net, dev) {
		int err;
		char fb_name[IFNAMSIZ];

		@@ -5083,7 +5084,7 @@ static void __net_exit default_device_exit(struct net *net)
		/* Delete virtual devices */
		if (dev->rtnl_link_ops && dev->rtnl_link_ops->dellink) {
		dev->rtnl_link_ops->dellink(dev);
		continue;
		goto restart;
		}

		/* Push remaing network devices to init_net */
		@@ -5094,6 +5095,7 @@ static void __net_exit default_device_exit(struct net *net)
		__func__, dev->name, err);
		BUG();
		}
		goto restart;
		}
		rtnl_unlock();
		}

Admin message